The 24th Conference on Uncertainty in Artificial Intelligence (UAI 2008)

The 21st Annual Conference on Learning Theory (COLT 2008)

Welcome to UAI and COLT! This sheet contains some practical information about these events. A separate sheet is available for ICML, MLG and the joint workshops. UAI and COLT take place in the Porthania building of University of Helsinki (Yliopistonkatu 3). ICML, MLG and the workshops take place earlier in the University Main Building (entrance at Fabianinkatu 33). Registration and information desk is located in the entrance hall of the Main Building on Wednesday (where the workshops and the University Reception take place) and in the entrance hall of the Porthania building on Thu-Sat. The opening times are as follows:

− Wed., July 9: 8:00-18:00 (Main Building) − Thu., July 10: 8:00-20:00 (Porthania) − Fri., July 11: 8:00-20:00 (Porthania) − Sat., July 12: 8:00-14:00 (Porthania)

Phone: 046 667 26 71 (from Finland) or +358 46 667 26 71 (international).

Lecture halls and other functions in Porthania are located as follows: − 1st (ground) floor: registration desk, lecture halls PI and PII − 2nd floor: room P219 (“lehtisali” i.e. “newspaper room”) − 7th floor: rooms P723 and P724

UAI sessions and all invited talks take place in PI. COLT sessions except for invited talks take place in PII. UAI and COLT delegates may freely switch between PI and PII, but you are encouraged to wait until session breaks to avoid hassle.

There is a short (5 min.) break after the joint invited talks in PI to allow COLT people to move to PII. Please make this switch as swift as possible to avoid delays in the program.

Coffee and UAI posters will be in the 1st (ground) floor lobby. Rooms P219, P723 and P724 are freely available for working on your laptop etc.

Lunches are on your own. There are several restaurants nearby, especially on Aleksanterinkatu and its side streets. University cafeterias in the main building and Porthania offer lunch at appr. EUR 5-7.

Proceedings for UAI or COLT are included in your registration package. All UAI, COLT and ICML papers are also freely available via the conference web pages. A limited number of copies of the paper proceedings are for sale at the registration desk. The university WLAN (HUPnet) can be accessed using your personal user id and password, printed on a piece of paper in your registration package. Outside the university, in the downtown area there are also several WLAN public hotspots available, some of which are completely free and some free for customers of cafeterias etc. (see http://ptp.hel.fi/wlan/start.asp?lang=3). Welcoming reception jointly for all participants of UAI, COLT, ICML, MLG and the workshops will be offered by the University of Helsinki on Wednesday, July 9, at 18:00-20:00 in the University Main Building (Fabianinkatu 33, 2nd floor). Banquet for UAI and COLT participants takes place in the White Hall (Valkoinen sali), Aleksanterinkatu 16-18, immediately after the last session on Friday. To get to the White Hall, exit Porthania and turn left. When you come to Senate Square, the White Hall is in the building diagonally across the square from you (“C” in the map below). Your name badge is your banquet ticket. If you ordered extra banquet tickets, they are in your registration package.  Public transportation: For getting around in the city, see the Journey Planner at http://aikataulut.ytv.fi/reittiopas/en/. Be warned that there are several active construction sites in the city, possibly affecting the traffic. Misc. information

− A university bookshop is located in Porthania − Photocopying machines operate with cards sold in the bookshop − Printing is available at Yliopistopaino (Vuorikatu 3, tel. 09 7010 2351) − Electricity has 230 V, 50 Hz (but availability in lecture halls is limited) − Helsinki City Tourist Office is located at Pohjoisesplanadi 19 − Taxi cab: tel. 0100 0700 − Emergency tel.: 112

 

UAI/COLT points of interest

A: Porthania, Yliopistonkatu 3: UAI, COLT sessions

B: Main Building, Fabianinkatu 33: ICML, MLG, workshops, welcoming reception

C: Valkoinen sali (White Hall), Aleksanterinkatu 16-18: joint UAI/COLT Banquet

D: Helsinki City Tourist Office, Pohjoisesplanadi 19

 

UAI/COLT SCHEDULE FOR THURSDAY, JULY 10, 2008 UAI COLT

8:30-8:35: COLT Opening Remarks 08:30

08:45 8:45-9:00 UAI Opening Remarks

Kamalika Chaudhuri and Satish Rao. Learning Mixtures of Product Distributions Using Correlations and Independence

08:35

09:00 David Mimno and Andrew McCallum. Topic Models Conditioned on Arbitrary Features with Dirichlet-multinomial Regression

Kamalika Chaudhuri and Satish Rao. Beyond Gaussians: Spectral Methods for Learning Mixtures of Heavy-Tailed Distributions

09:00

09:25 Amit Gruber and Michal Rosen Zvi, Yair Weiss. Latent Topic Models for Hypertext

Shai Ben-David, Tyler Lu and David Pal. Does Unlabeled Data Provably Help? Worst-case Analysis of the Sample Complexity of Semi-Supervised Learning

09:25

09:50

Topi

c M

odel

s an

d C

lust

erin

g

Daniel Tarlow, Richard Zemel and Brendan Frey. Flexible Priors for Exemplar-based Clustering

Maria-Florina Balcan, Steve Hanneke and Jennifer Wortman. The True Sample Complexity of Active Learning (Mark Fulk award)

Unsupervised, Sem

i-Supervised and A

ctive Learning

09:50

10:15 10:15-10:40: Coffee break

10:15

10:40 10:40-11:40: Invited talk by Peter Grünwald, CWI The Catch-Up Phenomenon in Bayesian Inference

10:40

11:40 11:40-11:45: Short break for switching rooms 11:40 11:45 Tamir Hazan and Amnon Shashua.

Convergent Message-Passing Algorithms for Inference over General Graphs with Convex Free Energies (Best student paper award runner up)

Elad Hazan and Satyen Kale. Extracting Certainty from Uncertainty: Regret Bounded by Variation in Costs

11:45

12:10 David Sontag, Talya Meltzer, Amir Globerson, Tommi Jaakkola and Yair Weiss. Tightening LP Relaxations for MAP using Message Passing (Best paper award)

Kosuke Ishibashi, Kohei Hatano and Masayuki Takeda. Online Learning of Maximum p-Norm Margin Classifiers with Bias

12:10

12:35

Infe

renc

e I

Justin Domke. Learning Convex Inference of Marginals

Subhash Khot and Ashok Kumar Ponnuswami. Minimizing Wide Range Regret with Time Selection Functions

On-Line Learning (I) 12:35

13:00 13:00-14:30: Lunch break

13:00

14:30 14:30-15:30: Invited talk by Robin Hanson, George Mason University Combinatorial Prediction Markets

14:30

15:30 15:30-15:35: Short break for switching rooms 15:30 15:35 Sevan Ficici, David Parkes and Avi Pfeffer.

Learning and Solving Many-Player Games through a Cluster-Based Representation

Nir Ailon and Mehryar Mohri. An Efficient Reduction of Ranking to Classification

15:35

16:00 Enrique Munoz de Cote and Michael Littman. A Polynomial-time Nash Equilibrium Algorithm for Repeated Stochastic Games (Best student paper award)

Michael Kearns and Jennifer Wortman. Learning from Collective Behavior

16:00

16:25 Gam

es a

nd

Dec

isio

ns

Cassio de Campos and Qiang Ji. Strategy Selection in Influence Diagrams using Imprecise Probabilities

Bharath Sriperumbudur, Arthur Gretton, Kenji Fukumizu, Gert Lanckriet and Bernhard Schölkopf. Injective Hilbert Space Embeddings of Probability Measures

Other D

irections (I)

16:25

16:50 16:50-17:15: Coffee break 16:50

Sung-Soon Choi, Kyomin Jung and Jeong Han Kim. Almost Tight Upper Bound for Finding Fourier Coefficients of Bounded Pseudo-Boolean Functions

17:15

Sandra Zilles, Steffen Lange, Robert Holte, and Martin Zinkevich. Teaching Dimensions based on Cooperative Learning

17:40

17:15

17:15-18:15: UAI Poster Spotlights

Vitaly Feldman. On the Power of Membership Queries in Agnostic Learning

Com

plexity and B

oolean Functions (I)

18:05

Vitaly Feldman and Leslie Valiant. The Learning Power of Evolution Parikshit Gopalan, Adam Kalai and Adam Klivans. A Query Algorithm for Agnostically Learning DNF? Adam Smith and Manfred Warmuth. Learning Rotations

Open

Problem

Session

18:30

19:00-20:00 Break (Refreshments at the UAI Poster Session)

19:00

18:15

18:15-21:00:

UAI Poster Session

21:00 20:00-21:30: COLT Business Meeting

20:00

UAI/COLT SCHEDULE FOR FRIDAY, JULY 11, 2008 UAI COLT

Thorsten Doliwa, Michael Kallweit and Hans Ulrich Simon. Dimension and Margin Bounds for Reflection-invariant Kernels

08:35

09:00 Kuzman Ganchev, Joao Graca, John Blitzer and Ben Taskar. Multi-View Learning over Structured and Non-Identical Outputs

Dana Angluin, James Aspnes, Jiang Chen, David Eisenstat and Lev Reyzin. Learning Acyclic Probabilistic Circuits Using Test Paths

09:00

09:25 C. Mario Christoudias, Raquel Urtasun and Trevor Darrell. Multi-View Learning in the Presence of View Disagreement

Linda Sellie. Learning Random Monotone DNF Under the Uniform Distribution

09:25

09:50

Mul

ti-Vi

ew a

nd T

rans

fer

Lear

ning

Gal Elidan, Benjamin Packer, Geremy Heitz and Daphne Koller. Convex Point Estimation using Undirected Bayesian Transfer Hierarchies

Eric Blais, Ryan O'Donnell and Karl Wimmer. Polynomial Regression under Arbitrary Product Distributions

Com

plexity and Boolean

Functions (II)

09:50

10:15 10:15-10:40: Coffee break

10:15

10:40 10:40-11:40: Invited talk by Gabor Lugosi, Pompeu Fabra University Concentration Inequalities

10:40

11:40 11:40-11:45: Short break for switching rooms 11:40 11:45 Emma Brunskill, Bethany Leffler, Lihong Li,

Michael Littman and Nicholas Roy. CORL: A Continuous-state Offset-dynamics Reinforcement Learner

Alon Zakai and Ya'acov Ritov. How Local Should a Learning Method Be?

11:45

12:10 Branislav Kveton and Milos Hauskrecht. Partitioned Linear Programming Approximations for MDPs

Yiming Ying and Colin Campbell. Learning Coordinate Gradients with Multi-Task Kernels

12:10

12:35

Rei

nfor

cem

ent l

earn

ing

Marc Toussaint, Laurent Charlin and Pascal Poupart. Hierarchical POMDP Controller Optimization by Likelihood Maximization (Best paper award runner up)

Vladimir Koltchinskii and Ming Yuan. Sparse Recovery in Large Ensembles of Kernel Machines

Generalization and

Statistics (I) 12:35

13:00 13:00-14:30: Lunch break

13:00

14:30 14:30-15:30: Invited talk by Dan Klein, UC Berkeley Unsupervised Learning for Natural Language Processing

14:30

15:30 15:30-15:35: Short break for switching rooms 15:30 15:35 Ulf Nielsen, Jean-Philippe Pellet and André

Elisseeff. Explanation Trees for Causal Bayesian Networks

Amy Greenwald, Zheng Li and Warren Schudy. More Efficient Internal-Regret-Minimizing Algorithms

15:35

16:00 Phil Dawid and Vanessa Didelez. Identifying Optimal Sequential Decisions

Giovanni Cavallanti, Nicolò Cesa-Bianchi and Claudio Gentile. Linear Algorithms for Online Multitask Classification

16:00

16:25

C

ausa

lity

Jin Tian. Identifying Dynamic Sequential Plans

Jacob Abernethy, Elad Hazan and Alexander Rakhlin. Competing in the Dark: An Efficient Algorithm for Bandit Linear Optimization (Machine Learning Journal award)

On-Line Learning and

Bandits

16:25

16:50 16:50-17:15: Coffee break 16:50

17:15 Harald Steck. Learning the Bayesian Network Structure: Dirichlet Prior vs. Data

Wouter Koolen and Steven de Rooij. Combining Expert Advice Efficiently

17:15

17:40 Gustavo Lacerda, Peter Spirtes, Joseph Ramsey and Patrik Hoyer. Discovering Cyclic Causal Models by Independent Components Analysis

Maria-Florina Balcan, Avrim Blum and Nathan Srebo. Improved Guarantees for Learning via Similarity Functions

17:40

18:05

Net

wor

k le

arni

ng

Varun Ganapathi, David Vickrey, John Duchi and Daphne Koller. Constrained Approximate Maximum Entropy Learning of Markov Random Fields

Benjamin I. P. Rubinstein and J. Hyam Rubinstein. Geometric & Topological Representations of Maximum Classes with Applications to Sample Compression

18:05

Shai Shalev-Shwartz and Yoram Singer. On the Equivalence of Weak Learnability and Linear Separability: New Relaxations and Efficient Boosting Algorithms

Other D

irections (II)

18:30 18:30

18:30.-19:55: UAI Business Meeting

18:55-19:55: COLT Rump Session 18:55

20:00 20:00-23:00: Joint UAI/COLT Banquet at Valkoinen Sali (the White Hall), address: Aleksanterinkatu 16-18

20:00

UAI/COLT SCHEDULE FOR SATURDAY, JULY 12, 2008 UAI COLT

Andrey Bernstein and Nahum Shimkin. Adaptive Aggregation for Reinforcement Learning with Efficient Exploration: Deterministic Domains

08:35

09:00 Keith Noto and Mark Craven. Learning Hidden Markov Models for Regression using Path Aggregation

Peter Bartlett, Varsha Dani, Thomas Hayes, Sham Kakade, Alexander Rakhlin and Ambuj Tewari. High-Probability Regret Bounds for Bandit Online Linear Optimization

09:00

09:25 Seyoung Kim and Eric Xing. Feature Selection via Block-Regularized Regression

Aleksandrs Slivkins and Eli Upfal. Adapting to a Changing Environment: the Brownian Restless Bandits

09:25

09:50 Mod

elin

g an

d R

egre

ssio

n

David Barber. Clique Matrices for Statistical Graph Decomposition and Parameterising Restricted Positive Definite Matrices

Varsha Dani, Thomas Hayes and Sham Kakade. Stochastic Linear Optimization under Bandit Feedback

Bandits and R

einforcement

Learning

09:50

10:15 10:15-10:40: Coffee break

10:15

10:40 Arthur Choi and Adnan Darwiche. Approximating the Partition Function by Deleting and then Correcting for Model Edges

Ohad Shamir and Naftali Tishby. Model Selection and Stability in k-means Clustering

10:40

11:05 Venkat Chandrasekaran, Nathan Srebro and Prahladh Harsha. Complexity of Inference in Graphical Models

Shai Ben-David and Ulrike von Luxburg. Relating Clustering Stability to Properties of Cluster Boundaries

11:05

11:30 Peter Thwaites, Jim Smith and Robert Cowell. Propagation using Chain Event Graphs

Kamalika Chaudhuri and Andrew McGregor. Finding Metric Structure in Information Theoretic Clustering

11:30

11:55

Infe

renc

e II

Tal El-Hay, Nir Friedman and Raz Kupferman. Gibbs Sampling in Factorized Continuous-Time Markov Processes (Best student paper award runner up)

Karthik Sridharan and Sham Kakade. An Information Theoretic Framework for Multi-view Learning

Unsupervised and Sem

i-Supervised Learning 11:55

12:20 12:20-13:50: Lunch break

12:20

13:50 Noah Goodman, Vikash Mansinghka, Daniel Roy, Keith Bonawitz and Joshua Tenenbaum. Church: a language for generative models

Jacob Abernethy, Peter Bartlett, Alexander Rakhlin and Ambuj Tewari. Optimal Stragies and Minimax Lower Bounds for Online Convex Games

13:50

14:15 Mathias Niepert, Dirk Van Gucht and Marc Gyssens. On the Conditional Independence Implication Problem: A Lattice-Theoretic Approach (Best student paper award runner up)

Robert Kleinberg, Alexandru Niculescu-Mizil and Yogeshwer Sharma. Regret Bounds for Sleeping Experts and Bandits (Machine Learning Journal award)

14:15

14:40 Hannaneh Hajishirzi and Eyal Amir: Sampling First Order Logical Particles

Jacob Abernethy, Manfred Warmuth and Joel Yellin. Optimal Strategies for Random Walks

14:40

15:05

Mod

elin

g

Jim Huang and Brendan Frey. Cumulative distribution networks and the derivative-sum-product algorithm (Best student paper award runner up)

András György, Gábor Lugosi and György Ottucsák. On-line Sequential Bin Packing

Online Learning (II) 15:05

15:30 15:30-15:55: Coffee break 15:30

Shuheng Zhou, John Lafferty and Larry Wasserman. Time Varying Undirected Graphs

15:55

Constantine Caramanis and Shie Mannor. Learning in the Limit with Adversarial Disturbances

16:20

Liwei Wang, Masashi Sugiyama, Cheng Yang, Zhi-Hua Zhou and Jufu Feng. On the Margin Explanation of Boosting Algorithms

16:45

Aarti Singh, Robert Nowack and Clayton Scott. Adaptive Hausdorff Estimation of Density Level Sets

17:10

Satyaki Mahalanabis and Daniel Stefankovic. Density estimation in linear time

Generalization and Statistics (II) 17:35

UAI Posters The poster session will take place on Thursday, July 10, at 18:15-21:00 in the lobby area of the Porthania building. The session is open for both UAI and COLT participants. Snacks available.

1 Umut A. Acar, Alexander Ihler, Ramgopal R. Mettu, Özgür Sumer Adaptive inference on general graphical models

2 Dimitrios Antos, Avi Pfeffer Identifying reasoning patterns in games

3 Vincent Auvray, Louis Wehenkel Learning Inclusion-Optimal Chordal Graphs

4 Debarun Bhattacharjya, Ross Shachter Sensitivity analysis in decision circuits

5 Liefeng Bo, Cristian Sminchisescu Greedy Block Coordinate Descent for Large Scale Gaussian Process Regression

6 Zhihong Cai, Manabu Kuroki On Identifying Total Effects in the Presence of Latent Variables and Selection bias

7 Botond Cseke, Tom Heskes Bounds on the Bethe Free Energy for Gaussian Networks

8 James Cussens Bayesian network learning by compiling to weighted MAX-SAT

9 Gert De Cooman, Filip Hermans, Erik Quaeghebeur Sensitivity analysis for finite Markov chains in discrete time

10 John Duchi, Stephen Gould, Daphne Koller Projected Subgradient Methods for Learning Sparse Gaussians

11 Quang Duong, Michael Wellman, Satinder Singh Knowledge Combination in Graphical Multiagent Models

12 Frederick Eberhardt Almost Optimal Intervention Sets for Causal Discovery

13 Vibhav Gogate, Rina Dechter AND/OR Importance Sampling

14 Peter Grünwald, Joe Halpern A Game-Theoretic Analysis of Updating Sets of Probabilities

15 Eric Hansen Sparse Stochastic Finite-State Controllers for POMDPs

16 Greg Hines, Kate Larson Learning When to Take Advice: A Statistical Test for Achieving A Correlated Equilibrium

17 Patrik Hoyer, Aapo Hyvärinen, Richard Scheines, Peter Spirtes, Joseph Ramsey, Gustavo Lacerda, Shohei Shimizu

Causal discovery of linear acyclic models with arbitrary distributions

18 Bowen Hui, Craig Boutilier Toward Experiential Utility Elicitation for Interface Customization

19 Alejandro Isaza, Csaba Szepesvari, Vadim Bulitko, Russell Greiner

Speeding Up Planning in Markov Decision Processes via Automatically Constructed Abstraction

20 Tony Jebara Bayesian Out-Trees

21 Manabu Kuroki, Zhihong Cai The Evaluation of Causal Effects in Studies with an Unobserved Exposure/Outcome Variable: Bounds and Identification

22 Johan Kwisthout, Linda van der Gaag The Computational Complexity of Sensitivity Analysis and Parameter Tuning

23 Eric Laber, Susan Murphy Small Sample Inference for Generalization Error in Classification Using the CUD Bound

24 Gregory Lawrence, Stuart Russell Improving Gradient Estimation by Incorporating Sensor Data

25 Daniel Lowd, Pedro Domingos Learning Arithmetic Circuits

26 Marina Meila, Le Bao Estimation and clustering with infinite rankings

27 Kurt Miller, Thomas Griffiths, Michael I. Jordan The Phylogenetic Indian Buffet Process: A Non-Exchangeable Nonparametric Prior for Latent Features

28 Lars Otten, Rina Dechter Bounding Search Space Size via (Hyper)tree Decompositions

29 Yan Radovilsky, Eyal Shimony Observation Subset Selection as Local Compilation of Performance Profiles

30 Sebastian Riedel Improving the Accuracy and Efficiency of MAP Inference for Markov Logic

31 Stephane Ross, Joelle Pineau Model-Based Bayesian Reinforcement Learning in Large Structured Domains

32 Aleksandr Simma, Moises Goldszmidt, John MacCormick, Paul Barham, Richard Black, Rebecca Isaacs, Richard Mortier

CT-NOR: Representing and Reasoning About Events in Continuous Time

33 Tomas Singliar, Denver Dash Efficient Inference in Persistent Dynamic Bayesian Networks

34 Matthew Streeter, Stephen Smith New Techniques for Algorithm Portfolio Design

35 Richard Sutton, Csaba Szepesvari, Alborz Geramifard, Michael Bowling

Dyna-Style Planning with Linear Function Approximation and Prioritized Sweeping

36 Jarno Vanhatalo, Aki Vehtari Modelling local and global phenomena with sparse Gaussian processes

37 Chong Wang, David Blei, David Heckerman Continuous Time Dynamic Topic Models

38 Max Welling, Yee Whye Teh, Bert Kappen Hybrid Variational/Gibbs Collapsed Inference in Topic Models

39 Ydo Wexler, Christopher Meek Inference for Multiplicative Models

40 Haohai Yu, Robert van Engelen Refractor Importance Sampling